Extracting Chisq values from logistf in R

Question

I am running a logistf model on R

fit<-logistf(data=data, Amylose_trans ~ Tissue + Species); summary(fit)
logistf(formula = Amylose_trans ~ Tissue + Species, data = data)

Model fitted by Penalized ML Confidence intervals and p-values by Profile Likelihood Profile Likelihood Profile Likelihood Profile Likelihood Profile Likelihood Profile Likelihood Profile Likelihood

                       coef se(coef) lower 0.95 upper 0.95       Chisq            p
(Intercept)      -2.9444390 1.529438  -7.803600 -0.8816003  9.89263874 1.659412e-03
TissueFresh_comb -0.8108826 2.128231  -6.917669  4.7311183  0.11508982 7.344222e-01
Tissueguts        6.7620487 2.026376   3.573546 12.5665756 29.41458648 5.843615e-08
TissueNodules    -1.0664308 2.053264  -6.764977  4.2922631  0.23665005 6.266362e-01
TissueOld_comb   -1.0664308 2.053264  -6.764977  4.2922631  0.23665005 6.266362e-01
SpeciesO.badius   0.2386536 1.461843  -5.491559  5.5635990  0.01074908 9.174251e-01
SpeciesO.sp.      0.2386536 1.461843  -5.491559  5.5635990  0.01074908 9.174251e-01

Likelihood ratio test=140.6088 on 6 df, p=0, n=144
Wald test = 31.04725 on 6 df, p = 2.482811e-05

Covariance-Matrix:

              [,1]      [,2]       [,3]       [,4]       [,5]          [,6]          [,7]
[1,]  2.339181e+00 -2.339181 -2.3391813 -2.3391813 -2.3391813  1.421471e-15  1.178081e-15
[2,] -2.339181e+00  4.529369  2.8174918  2.9407710  2.9407710 -1.080011e+00 -1.080011e+00
[3,] -2.339181e+00  2.817492  4.1061987  2.7118392  2.7118392 -6.690186e-01 -6.690186e-01
[4,] -2.339181e+00  2.940771  2.7118392  4.2158934  2.8078876 -8.414506e-01 -8.414506e-01
[5,] -2.339181e+00  2.940771  2.7118392  2.8078876  4.2158934 -8.414506e-01 -8.414506e-01
[6,]  1.421471e-15 -1.080011 -0.6690186 -0.8414506 -0.8414506  2.136985e+00  8.842641e-01
[7,]  1.178081e-15 -1.080011 -0.6690186 -0.8414506 -0.8414506  8.842641e-01  2.136985e+00

and I have been trying to extract the Chisq values but I cannot figure out how to do it! I can easily extract the coef, terms, p.values, lower and upper ci etc etc but I cant figure out a way to extract the Chisq values

any suggestions?

thanks in advance

Hey I am not sure what you are talking about, I checked it which is where I was looking in the first place! — Panos, Feb 04 '17 at 10:06
ok, it is more difficult than I first supposed. I found in the source of the function `logistf()` this part: `fit$prob[i] <- 1 - pchisq(2 * (fit.full$loglik - fit.i$loglik), 1)` Type `logistf` on the command line and enter to see the definition of the function. — jogo, Feb 05 '17 at 10:15

score 0 · Answer 1 · edited Mar 17 '21 at 20:52

0

The following should do the trick:

qchisq(1-fit$prob,df=1)

edited Mar 17 '21 at 20:52

Tomerikoo

18,379
16
47
61

answered Mar 15 '21 at 14:33

rskim

1

score -1 · Answer 2 · answered Dec 17 '20 at 22:55

-1

"drop1(fit)" This should work.

answered Dec 17 '20 at 22:55

user14847103

1

2

This doesn't seem to answer the question. Can you elaborate? – kangaroo_cliff Dec 17 '20 at 22:58
how does this answer the question? – StupidWolf Dec 18 '20 at 03:07

Extracting Chisq values from logistf in R

2 Answers2